A study of the use of spell and grammar checker in texts by second foreign language learners of Spanish. Henrietta Carolsson Godolakis, Spanish department at the University of Lund
_____________________________________________________________________________
Abstract
This study was made in order to evaluate the didactic use of the spell and grammar checker in texts by
second foreign language learners of Spanish at an upper secondary school in Sweden. Four students
participated in the study.
It was shown in the study that Microsoft Word 2010 detected about 40% of the mistakes found in the
students’ texts. Out of the by MS Word detected mistakes, the programme gave the correct feedback on 69%.
Thus, it is concluded that MS Word correctly detected, i.e. detected and provided the correct feedback, to
28% of the mistakes made by the students. Furthermore, it is understood that there was a difference between
the efficiency of MS Word when detecting spelling and grammar mistakes. 59% of all the spelling mistakes
found in the students’ texts was detected correctly by MS Word, while no more than 15% of the grammar
mistakes were correctly detected.
The article gives an overview on how efficiently MS Word detected different types of mistakes concerning
spelling as well as grammar. Furthermore, an analysis on how the students used the feedback by MS Word is
provided.
It is concluded in the article that the spell and grammar checker provided by Microsoft Word 2010 is not
alone a sufficient tool for a second foreign language learner of Spanish in order to improve their written
texts, but could be used in order to improve a text superficially.
Keywords:
Spanish, foreign language, spelling, grammar, mistakes, Microsoft Word,
______________________________________________________________________________
student
1. Introduction
Apart from Swedish and English, all Swedish students have the opportunity to study at least one more
foreign language in school, Spanish being one of those. As expressed in the curriculum of languages, the
students should develop comprehensive communicative skills. This involves receptive skills, as well as
productive and interactive skills. Furthermore, the students should be given the opportunity to develop
strategies in order to support their communication and to solve problems caused by their limited proficiency.
The teaching should enable the students to speak, interact and write with the support of different means and
media1.
One example of how students could be supported in the development of their language is to use the spell and
grammar programme provided by word processors. Today a considerable amount of all written production in
school is made on a computer. This provides the students the possibility to use different programmes to
check spelling and grammar when revising a text. Microsoft Word is a widely used programme on computers
today and for students, it is a commonly used tool. However, little is known about the didactic capacity of
the tool in foreign language learning. Thus, this study was designed to study how the students use the
feedback provided by the programme in order to process their texts.
2. Background A couple of studies have been made concerning learner responses on corrective feedback and the use of a
spell and grammar checker. Trude Heift, professor at the Linguistics department of Simon Fraser University,
conducted a study in order to find out the response on corrective feedback from a computer programme of
Canadian university students. It was concluded that the students responded more correctly when given
feedback that provided an explanation of the error and that also highlighted the error. Proficiency level and
gender did not affect the results2. The following year Anne Rimrott, linguist at the Linguistics department of
Simon Fraser University, and Trude Heift made an evaluation of the performance of the programme MS
Word using learners of German. The investigators examined spelling mistakes made by learners of German
and reached the conclusion that 80% of the spelling mistakes were competence errors rather than
typographical mistakes, i.e. performance errors. It was also concluded that the spell checker failed “to detect
or provide a correction for 48% of the spelling mistakes” made by the students3. Another study was made by
Heift and Rimrott in 2008 that was meant to investigate the “learner responses to three distinct types of
corrective feedback for misspellings produced by English learners of German while using the E-tutor, a
parser-based online CALL program”4. It was concluded in this study that the students submitted the target
word more often when given as an option in the feedback provided by the programme. Also it was suggested
that the order in which the words appeared influenced the students’ choice.
3. This study The material for this article was gathered at an upper secondary school in Sweden in February, 2013. A total
of 314 errors from texts made by four students of Spanish of the second grade were analysed.
1
http://www.skolverket.se/forskola-­‐och-­‐skola/gymnasieutbildning/amnes-­‐och-­‐laroplaner/mod Heift, T. (2004) Corrective feedback and learner uptake in CALL. . ReCALL 16 (2). Cambridge University Press, p 413-­‐
431. 3
Rimrott, A. & Heift, (2005) T. Language Learners and Generic Spell Checkers in CALL. CALICO Journal, Vol. 23, No. 1, p 17-­‐48. 4
Heift, T. & Rimrott, A. (2008) Learner responses to corrective feedback for spelling errors in CALL. System 36, p 196–
213. 2
3.1. Study participants The four students studied a course of Spanish equivalent to B1.15 according to the Common European
Framework of Reference for Languages (CEFR)6. They all had studied Spanish for four years in school and
one and a half year at the upper secondary school where they all attended the programme of social science.
All informants were female turning 18 this year. The informants’ proficiency level was measured by the
grade they received in June 2012 when they finished their last course of Spanish7. The students participating
in the study had received the following grades at the end of their last Spanish course:
Table 1. INFORMATION ABOUT INFORMANTS8.
Informant
Grade
Length of text (words)
Ana
B
344
Berta
A
381
Cecilia
B
337
Dora
F
208
Dora was attending course 4 of Spanish whilst studying extra to be able to pass the previous course. The
texts the students produced were of different length and of different proficiency level. Ana wrote her text in
the past tense as well as Berta. Both Dora’s and Cecilia’s texts were written in the present tense.
3.2. Procedures The students were given the instructions to write a text in Spanish describing a journey to Italy, as shown in a
series of pictures made by Malin Ågren9, to a person who could not see the pictures. They were advised to
write something about each picture and additionally they were asked to be detailed in their descriptions.
They had no time limit in order to complete their texts. However, the students used between 40 and 50
minutes to write. The programme used when writing had no spell or grammar check and the students could
not use any other tools when writing. All the students wrote their texts simultaneously. The texts were saved
in their original version and thereafter, the texts were copied into an MS Word document in order to use the
spell and grammar checker of this programme.
The students were then instructed to revise their texts using the spell and grammar checker. They were made
aware of the fact that MS Word does not always give the correct answer. Also they were instructed on how to
use the programme and that they could work in whatever order they liked. All students were already familiar
with the spell and grammar checker saying that they used it frequently when writing in English and to some
extent in Spanish as well. Ana and Berta worked through their texts the same day, while Cecilia and Dora did
it the following day. All students worked through their texts by clicking on the marked words in the text, thus
not receiving the extra feedback given by MS Words.
When revising their texts, the editing process was recorded by a programme called Screen-cast-o-matic10.
When their work was finished, the film was saved. Directly after the revision of the texts the students took
part in a stimulated recall, an introspective method suitable for examining processes. The method is often
used to study learning processes11. The student was asked to comment on her errors, the response from MS
Word on her mistakes and other interesting observations made on the film. The introspection was audio
recorded. One problem occurred during the stimulated recall of Ana which resulted in the fact that no
5
In Sweden you can study all together seven different courses during the second part of the 9-­‐year comprehensive school and upper secondary school. 6
http://www.coe.int/t/dg4/linguistic/CADRE1_EN.asp 7
The Swedish grades are A, B, C, D, E and F. An F indicates that you do not have sufficient knowledge to pass the course. 8
The students’ names have been changed. 9
Malin Ågren, 2008. Appendix 1. 10
This programme is a web based recording programme used for free. http://www.screencast-­‐o-­‐matic.com/ 11
http://srmo.sagepub.com/view/the-­‐sage-­‐dictionary-­‐of-­‐qualitative-­‐management-­‐research/n100.xml recording was made of the interview. The investigator noted this directly and documented the interview
immediately. Finally, the students were asked to grade the spell and grammar checker according to a Likertscale12.
The material was finally coded according to the type of error committed, the possible detection of the
mistakes, the possible response given by MS Word and also the response the students gave to the feedback
provided my MS Word.
3.3. Classification of Errors In order to be able to code the mistakes made by the students, the definition of a spelling error used by
Rimrott & Heift was used. According to them, a misspelling is “a nonexistent word in a given language
independent of the source of the error”13. The classification of errors used in this study was influenced by
Rimrott and Heift when it comes to classifying the spelling errors by performance and competence errors,
but also the subdivisions within these two. MS Word was originally developed in order to detect performance
errors made by adult native writers, but a non-native writer makes more competence errors. In their study,
Rimrott and Heift concluded that MS Word has difficulties in finding competence errors14.
Table 2. CLASSIFICATION OF PERFORMANCE ERRORS. Examples.
Addition
*receptción/recepción
Omission
*vistaron/visitaron
Single letter violations
Substitution
*numero/número
Transposition
*pusieorn/pusieron
Multiple letter violations
*reseptionista/recepcionista
Word boundary violations
*largo.Lisa
Adapted from Rimrott & Heift
(2005)
The examples above explain rather explicitly the different subdivisions within the category of performance
errors. The last example shows one of the few mistakes made concerning word boundary violations, where
the student’s performance error violates the word boundary of the intended words.
The codes used within the category of competence errors were interlingual, intralingual or wrong word
errors. The Unclassified –category was removed from the original coding schedule since there were no
examples put into this category. Interlingual errors were mistakes influenced by the students’ mother tongue
or possibly English15. Intralingual errors were mistakes made by the influence of the target language. The
Wrong word error was used when the student had produced a misspelt word that was an incorrect word in
the context.
Table 3.
Examples.
CLASSIFICATION
Language influence
Interlingual
*gitarra/guitarra
*resturante/restaurante
Adapted from Rimrott & Heift
(2005)
OF
COMPETENCE
Intralingual
*juegó/jugó
*granda/grande
ERRORS.
Wrong word
*de mas/ demasiado
*camninas/camas
12
Appendix 2. Rimrott, A. & Heift, T. (2005) Language Learners and Generic Spell Checkers in CALL. CALICO Journal, Vol. 23, No. 1, p 21. 14
Rimrott, A. & Heift, (2005) T. Language Learners and Generic Spell Checkers in CALL. CALICO Journal, Vol. 23, No. 1. 15
English is studied by Swedish children from an early age. Furthermore, the input of English is immense in Swedish society. 13
Additionally, the category of grammar was added to the study in order to better suit the learners of a second
foreign language, who are more likely to make grammar mistakes. To this category were added the mistakes
that were not coded as a spelling mistakes, according to the earlier definition. Codes used here were
agreement, grammatical word, verb form/tense, verb form/wrong verb, verb form and other. Agreement
contains mistakes in general concerning adjectives, verbs and articles that do not concord with their context.
Grammatical words are for example wrong choice of preposition, missing preposition or wrong usage of part
of speech. Swedish learners of Spanish generally make many verb mistakes since the verb system differ
largely from their mother tongue. Hence the verb form category was divided into three. When the student
chose the wrong tense, the mistake was coded as Verb form/tense. Verb form/wrong verb was used when the
student had inflected the verb correctly but had used the wrong verb in the context. These mistakes were
included in the study even if they were of a more semantic character. When the student used an inflected
form of a verb that should be in the infinitive, or the opposite, verb form was used to code the mistake. When
a student used a word that did not work in the context, the code other was used. This category was also used
when the student repeated the same word twice.
Table 4. CLASSIFICATION OF GRAMMATICAL ERRORS.
Examples.
Agreement
Grammatical word
Verb form/Tense
Verb form/Wrong verb
Verb form
Other
un/una; rubio/rubia; amo/ama
por/para; -/de; delante/encima
van/fueron; pueden/podían
había/hacía; es/hace
va/ir
mira/vista; repetition of a word;
3.4. Research questions The main object of this study was to explore the potentials of Microsoft Word 2010, a tool originally
designed for native speakers, within the foreign language education area. What mistakes does the programme
detect and does it give correct feedback on these mistakes? Furthermore, how do the students use of the
feedback given by the spell and grammar checker? Can the tool be seen as beneficial for students learning
Spanish as a second foreign language? The aspect of proficiency was taken into consideration when
analyzing the material.
4. Results Firstly, the results of the errors coded as spelling mistakes and the grammatical mistakes will be analysed.
Secondly, an analysis of the learner responses will be made and, finally, the results of the stimulated recall.
4.1. Evaluation of spell and grammar checker In the study there were 314 mistakes made by the students. Out of these mistakes there were 91 spelling
mistakes and 223 grammar mistakes, i.e. 29% of the mistakes were spelling mistakes and 71% grammar
mistakes. Naturally, there were individual differences between the students depending on their level of
proficiency in their texts. Roughly speaking, one third of the students’ mistakes were spelling mistakes and
two thirds were grammatical errors.
Out of the 314 mistakes, 127 were detected by MS Word, i.e. approximately 40% of the mistakes. Out of the
detected errors, 88 were correctly detected, i.e. 69% of the detected errors. By correctly detected it is meant
that MS Word supplied the correct feedback. It is interesting to point out that only 88 out of 314 mistakes
were correctly detected, i.e. 28% of the mistakes made were correctly detected. In only one case did Word
flag a mistake that was not a mistake, i.e. a false alarm.
Table 5. STUDENT ERRORS.
Student
errors
Subject
Domain
Total
Ana
Spelling
Grammar
Spelling
Grammar
Spelling
Grammar
Spelling
Grammar
Total
16
81
16
34
28
48
31
60
314
Berta
Cecilia
Dora
Total
Occs
15
10
15
18
25
7
31
6
127
Errors detected by Word
Correctly
detected
Perc
Occs
Perc
94%
11
73%
12%
7
70%
94%
10
67%
53%
15
83%
89%
18
72%
15%
7
100%
100%
15
48%
10%
5
83%
40%
88
69%
False
alarms
Occs
0
0
0
1
0
0
0
0
1
Perc
0%
0%
0%
6%
0%
0%
0%
0%
1%
4.1.1 Spelling mistakes Examining the spelling mistakes made by the students, we see that 91 mistakes were made. 86 of these were
detected, i.e. almost 95%. However, only 54 mistakes were successfully corrected by MS Word, i.e. 63% of
the detected mistakes and 59% of the total amount of spelling mistakes.
Table
6.
RRORS.
SPELLING
Student
errors
Subject Domain Total
Ana
Berta
Cecilia
Dora
Spelling
Spelling
Spelling
Spelling
Total
16
16
28
31
91
Total
Occs
15
15
25
31
86
Errors detected by Word
Correctly
detected
Perc
Occs
Perc
94%
11
73%
94%
10
67%
89%
18
72%
100%
15
48%
95%
54
63%
False
alarms
Occs
0
0
0
0
0
Perc
0%
0%
0%
0%
0%
Examining the difference between competence and performance errors we can conclude that MS Word more
correctly detects performance errors, as concluded by Rimrott and Heift as well16. 50 performance errors
were made, 47 of these were found, i.e. 94%. Out of the detected performance mistakes, 42 were correctly
detected, 89%. This means that 84% of all performance errors were correctly detected.
Table 7. PERFORMANCE ERRORS.
16
Rimrott, A. & Heift, (2005) T. Language Learners and Generic Spell Checkers in CALL. CALICO Journal, Vol. 23, No. 1. Subject
Ana
Berta
Cecilia
Dora
Domain
Performance
errors
Performance
errors
Performance
errors
Performance
errors
Corrected
Uncorrected
Undetected
Total
Occs
Perc
Occs
Perc
Occs
Perc
Occs Perc
8
89%
0
0%
1
11% 9
100%
5
100%
0
0%
0
0%
100%
16
80%
2
10%
2
10% 20
100%
13
81%
3
19%
0
0%
16
100%
84%
5
10%
3
6%
50
100%
Total
42
Adapted from Rimrott & Heift (2005)
5
When analyzing the performance errors the majority of the students’ mistakes were coded as substitution, i.e.
58% of the mistakes. This is explained by the fact that errors where students had substituted a letter with an
apostrophe with a regular letter, or the opposite, the error was coded within this category. Omission counted
for 14% of the mistakes.
Table 8. CLASSIFICATION OF PERFORMANCE ERRORS.
Ana
Occs
Addition
1
Single
Omission
1
letter
Substitution
5
violations
Transposition 0
Multiple
letter
violations
0
Word
boundary
violations
2
Other
0
Total
9
Adapted from Rimrott & Heift (2005)
%
11%
11%
56%
0%
Berta
Occs
0
2
2
1
0%
0
22% 0
0%
0
100% 5
%
0%
40%
40%
20%
Cecilia
Occs %
4
20%
1
5%
11
55%
3
15%
Dora
Occs
0
3
11
1
0%
0
1
0%
1
0%
0
100% 20
0%
5%
0
0%
0
100% 16
%
0%
19%
69%
6%
Total
Occs
5
7
29
5
%
10%
14%
58%
10%
6%
1
2%
0%
3
0%
0
100% 50
6%
0%
100%
All in all, 41 competence errors were made, out of which 39 were detected, 95%. Only 12 of them were
correctly detected, i.e. 29%. It was concluded that 66% of the competence errors detected were left
uncorrected.
Table 9. COMPETENCE ERRORS
Subject
Domain
Corrected
Occs Perc
Competence
Ana
errors
3
Competence
Berta
errors
5
Competence
Cecilia
errors
2
Competence
Dora
errors
2
Total
12
Adapted from Rimrott & Heift (2005)
Uncorrected
Occs Perc
Undetected
Occs
Perc
Total
Occs
Perc
43%
4
57%
0
0%
7
100%
45%
5
45%
1
1%
11
100%
25%
5
63%
1
13%
8
100%
13%
29%
13
27
87%
66%
0
2
0%
5%
15
41
100%
100%
Looking closer at the different types of competence errors, we can conclude that close to two thirds of the
errors were due to intralingual interference and almost one third was due to interlingual interference from
mother tongue. These findings confirm the results found by Rimrott and Heift17.
Table 10. CLASSIFICATION OF COMPETENCE ERRORS.
Language influence
Subject
Ana
Berta
Cecilia
Dora
Total
Adapted
from Rimrott
&
Heift
(2005)
Interlingual
Occs
%
1
14%
3
27%
3
38%
4
27%
11
27%
Intralingual
Occs
%
6
86%
7
64%
4
50%
9
60%
26
63%
Wrong
word
Occs %
0
0
1
9
1
12
2
13
4
10%
Total
Occs
7
11
8
15
41
%
100%
100%
100%
100%
100%
4.1.2. Grammar mistakes Out of the 314 mistakes in the informants’ texts, there were 223 grammar mistakes. Out of the 223 grammar
mistakes only 41 were detected by Word, 18%. Out of the 41 detected mistakes 34 were correctly detected,
i.e. 83%. This indicates that only about 15% of the grammar mistakes committed in the informants’ texts
were actually corrected successfully by MS Word. It is obvious that the programme has difficulties in
detecting all kinds of grammatical errors that the students make. One mistake detected by MS Words was a
false alarm.
Table
11.
GRAMMAR
17
Rimrott, A. & Heift, (2005) T. Language Learners and Generic Spell Checkers in CALL. CALICO Journal, Vol. 23, No. 1. MISTAKES
Student
errors
Ana
Berta
Cecilia
Dora
Domain
Total
Grammar
Grammar
Grammar
Grammar
Total
81
34
48
60
223
Errors detected by Word
Correctly
detected
Perc
Occ
Perc
12%
7
70%
53%
15
83%
15%
7
100%
10%
5
83%
18%
34
83%
Total
Occ
10
18
7
6
41
False
alarms
Occ
0
1
0
0
1
Perc
0%
6%
0%
0%
2%
As shown in the table below, most of the students’ grammar mistakes were mistakes concerning agreement,
37%. The grammatical word mistakes and those of verb form/tense roughly represented one quarter of the
mistakes each. For a Swedish learner, this is not a surprising result.
Table 12. GRAMMAR
RESULTS
Agreement
Word order
Grammatical
word
Verb form/
Tense
Verb form/
Wrong Verb
Verb form
Other
Total
Corrected
Occs
Perc
29
85%
0
0%
Uncorrected
Occs Perc
0
0%
0
0%
Undetected
Occs
Perc
54
30%
0
0%
Total
Occs
83
0
Perc
37%
0%
4
12%
1
17%
54
30%
59
26%
0
0%
4
66%
51
28%
55
25%
0
0
1
34
0%
0%
3%
15%
0
1
0
6
0%
17%
0%
3%
16
4
4
183
9%
2%
2%
82%
16
5
5
223
7%
2%
2%
100%
From the table above, it is concluded that 85% of the by MS Word correctly detected mistakes was
agreement errors. However, studying these mistakes closer it is concluded that MS Word was unable to
correct the agreement errors concerning verbs. When analyzing the undetected words it is seen that three
categories of mistakes, agreement, grammatical word and verb form/tense, each counted for almost a third of
the mistakes.
Looking at the individual results we can see that Berta stood out in contrast to the others. 44% of her
grammar mistakes were successfully detected by MS Word. Compare this to the 15% of the overall result.
Almost 40% of her grammatical mistakes were agreement errors and all of these were detected. She also
made fewer grammar mistakes compared to the other. This partly explained her differentiated results.
4.2. Learner responses to correction suggestions When the students revised their texts they used MS Word 2010 spell and grammar checker. The feedback
provided by MS Word was coded, as well as the students’ response, and then analysed.
4.2.1. Spelling mistakes In total, the informants made 91 spelling mistakes, 88 were correctly detected. In 78 of the cases the students
made a correction taken from the feedback provided by MS Word. In 50 of the cases the student chose the
target word, i.e. 64% of the cases. In 47 of these cases the target word was found in the first position of the
list provided by MS Word. In 28 of the cases the students chose an incorrect word from the list provided by
MS Word. In 19 of these cases the students chose the first word of the list, 68%. This indicates a general
tendency among the students to choose the first word provided by MS Word. In 85% of the cases the students
chose a word from the first position of the list provided by MS Word.
Tabell 13. LEARNER RESPONSES TO FEEDBACK ON
SPELL-CHECKING ERRORS
Subject Action
Ana
Berta
Cecilia
Dora
Submits target
word
Submits
a
wrong
word
that is in the
list
Submits target
word
Submits
a
wrong
word
that is in the
list
Submits target
word
Submits
a
wrong
word
that is in the
list
Submits target
word
Submits
a
wrong
word
that is in the
list
Total
Adapted from
Heift
&
Rimrott (2008)
Position of selected word
1st
2nd
3rd
4th
>4th
Occs %
Occs %
Occs %
Occs %
9
90%
0
0%
Occs %
10
1
%
Total
Occ
s
%
0
0%
0
0%
10
100%
2
0
0%
1
20
%
2
40
%
0
0%
5
100%
9
40%
100
%
0
0%
0
0%
0
0%
0
0%
9
100%
1
50%
0
0%
0
0%
0
0%
1
50
%
2
100%
16
94%
1
6%
0
0%
0
0%
0
0%
17
100%
5
100
%
0
0%
0
0%
0
0%
0
0%
5
100%
13
93%
1
7%
0
0%
0
0%
0
0%
14
100%
11
66
69% 4
85% 6
25
%
8%
0
2
0%
3%
1
3
6%
4%
0
1
0%
1%
16
78
100%
100%
4.2.2. Grammar mistakes All in all the students made 223 grammar mistakes but only 40 of them were detected. Out of the 40 detected
mistakes, 34 were correctly detected, 85%. In 32 of the cases an appropriate change was made by the
student, 94%. MS Word is rather successful in giving the correct feedback in these cases and the students
made an appropriate change given the feedback from the programme. Many of the correctly detected
mistakes were mistakes concerning agreement, mistakes that are easily understood by the students. In the 6
cases where MS Word found the mistake but failed to give the target word, in 15% of the cases, the students
made an inappropriate change each time, something that indicated that the students relied on the feedback
given MS Word.
4.3. Learner comments in stimulated recall session In this part a few interesting examples of how the student worked or reacted to the feedback that MS Word
gave them will be presented. The results will be presented individually. This part of the article ends with the
students’ evaluation of the spell and grammar checker by MS Word.
4.3.1. Ana As mentioned earlier Ana’s retrospection was never recorded but the investigator wrote down the comments
made by her directly after the interview.
Ana was very conscious when processing her text. She read carefully through her text when working with
the spell and grammar check. She pointed out that MS Word was not suitable for correcting verb forms
though. Still she chose the incorrect verb form from the list provided by MS Word a couple of times. She
took time to really think about what she wanted to express before she made any corrections. Ana finished her
work by reading through her text dividing it into paragraphs.
4.3.2 Berta Berta made few mistakes in general and worked rapidly through her text. When doing so she was reading her
text though. A couple of times she pointed out that MS Word did not provide any good alternatives at all and
that she trusted herself even more. At one time MS Word indicated an error that was actually a false alarm,
and Berta changed her form according to the feedback given. When finished processing the text, she went
back and had a second look at it. She commented on having changed an incorrectly inflected verb into
another verb that she knew how to inflect in order to please MS Word.
4.3.3. Cecilia Cecilia worked through her text quickly. The performance errors she had made were successfully detected
and she corrected them easily. She stopped for a longer time twice to think about the feedback given by MS
Word. Once when she had to choose a verb tense, and once when she had made up a word that MS Word did
not recognise. When finished revising her text she had a second look at some mistakes that had occurred
after making some corrections.
4.3.4. Dora Dora worked rapidly through her text being used to the tool provided by MS Word. She easily corrected the
performance errors since she recognised the word directly from the feedback given. Sometimes when she did
not know a word, she had invented one from her previous knowledge of Spanish or English. When MS Word
was unable to guess the target word, Dora just chose one from the list that she thought would do, something
that at times made the sentence more confusing. Dora focused only on the feedback from MS Word. At one
time she went back in the text to see what she had written in the sentence before. Her revising of the text
lacked reflections to what MS Word meant by its feedback.
4.3.5. Likert-­‐scale. At the end of the retrospection the students were supposed to evaluate the spell and grammar check provided
by MS Word according to the Likert-scale. The students’ evaluation is summarised in the following table:
Table 14. EVALUATION OF SPELL AND GRAMMAR
CHECKER.
Likert scale
Informant
Ana
Berta
Cecilia
Dora
Grade
3-4
3
4
3
Ana graded the programme by a grade between 3 and 4 explaining that it worked rather well with simple
spelling and agreement mistakes, but that it had its flaws when it was dealing with grammar mistakes. Berta
affirmed that MS Word helped her out, reminding her about the masculine form of certain nouns. Using the
Likert-scale she evaluated the programme by giving it grade 3. Cecilia did not feel that she had learnt
anything from the use of the feedback provided by MS Word. She was used to working with the programme
and thought that it was suitable for correcting spelling mistakes, referring at the types of mistakes coded as
performance errors in this study. She thought that MS Word was good, that it found the mistakes you tend to
make and also correct them. Using the Likert-scale Cecilia gave the programme a 4. Dora rejected having
learnt anything from working with the feedback given from MS Word saying that she did not trust the
programme since it often failed to give the correct feedback. She also explained that the tool was more useful
when she wrote in English. She makes fewer mistakes in English compared to Spanish. When writing in
Spanish she frequently looks the word up on the internet when she is uncertain of how to spell it. In Dora’s
final evaluation she gave the programme a grade 3 on the Likert-scale, saying that she could not rely on it to
help her out.
5. Conclusion and further suggestions and research From the results of this study it is concluded that MS Word efficiently detects all types of spelling errors
made by students of Spanish as a second foreign language, and that the programme successfully gives correct
feedback when the errors are so called performance errors. However, the programme is not efficient enough
in giving correct feedback on the competence errors. Furthermore, the programme fails when it comes to
detecting grammar errors. The programme provides a high amount of correct feedback on the grammar
errors that it does detect, though. These grammar mistakes mostly concern agreement, excluding verb
agreement though. It is concluded that MS Word fails to detect and correct verb errors, in general. Since
verbs are considered one of the most important grammatical features in Spanish in order to communicate
successfully, this is alarming.
The proficiency of the students seems to affect the success of the programme, since a more proficient user
makes fewer mistakes, or of a more simple character, and thus these errors are more easily detected by MS
Word. More complicated grammatical errors go undetected. According to this study, the beneficial use of the
programme increase with the proficiency of the students both when it comes to errors detected and how to
use the feedback. The students in general have an excessive belief in the ability of the programme even
though they state that they are critical in their use of it.
To answer the general question of this study, it is concluded that the tool can be used beneficially by a
student learning Spanish as a second foreign language when it comes to polishing up their texts. The didactic
implications of this study would be to improve students’ own ability to detect their own mistakes, e. g. the
students should revise their own texts looking at for example agreement and verbs, these being their most
frequent mistakes. Then, after scrutinising their texts, they should be taught how to use the programme more
efficiently, using all feedback given by MS Word. However, when in need of a more thorough revision of a
text, a teacher is more efficient than MS Word.
In order for a second language learner to be able to rely more on MS Word, it is needed to develop it in order
for it to detect more mistakes correctly, especially when it comes to grammar mistakes, and improve the
programme in order for it to increase the correct feedback on for example the competence errors that it
detects but is unable to correct today. It would also be interesting to conduct further investigation on a larger
scale, thus receiving a more complete picture of the mistakes made by second language learners and the
mistakes detected and corrected by the programme. Additionally, it would be interesting to investigate if
working with MS Word can actually develop the students’ grammatical awareness in or if it limits the
students’ way of expressing themselves in order to please the programme.
References Council of Europe. (2013) Common European Framework of Reference for Languages: Learning, teaching,
assessment (CEFR) at Council of Europe [on-line]. Available at:
http://www.coe.int/t/dg4/linguistic/Source/Framework_en.pdf [2013-04-07]
Heift, T. (2004) Corrective feedback and learner uptake in CALL. ReCALL 16 (2). Cambridge University
Press, p 413-431.
Heift, T. & Rimrott, A. (2008). Learner Responses to Corrective Feedback for Spelling Errors in CALL.
System, 36(2), pp. 196-213.
Knutsson O., Ceratto Pargman T., Severinsson Eklund K. (2003) Transforming Grammar Checking
Technology into a Learning Environment for Second Language Writing. p 38-45
Rimrott, A. & Heift, T. (2005) Language Learners and Generic Spell Checkers in CALL. CALICO Journal,
Vol. 23, No. 1, p 17-48.
The SAGE Dictionary of Qualitative Management Research. [on-line] Available at:
http://srmo.sagepub.com/view/the-sage-dictionary-of-qualitative-management-research/n100.xml [2013-0308]
Screen-cast-o-matic. Available at: http://www.screencast-o-matic.com/ [2013-03-08]
Skolverket. (2011) “Ämne-Moderna språk” at Skolverket [on-line]. Available at:
http://www.skolverket.se/forskola-och-skola/gymnasieutbildning/amnes-och-laroplaner/mod [2013-03-17]
Ågren, M. (2008) À la recherche de la morphologie silencieuse : sur le développement du pluriel en francais
L2 écrit (Le voyage en Italie); Lunds universitet.
Appendix 1 El viaje a Italia
Appendix 2 The Likert-scale
Knutsson O., Ceratto Pargman T., Severinsson Eklund, K. (2003).
Grade 5
Grade 4
Grade 3
Excellent
Good
Acceptable
Grade 2
Bad
Grade 1
Incomprehensible
-I understand exactly what Word suggests.
-Word is a quite good help for me.
-It is hard for me to make up my mind on what Word says, but I
take a chance that Word is right.
-It is hard for me to make up my mind on what Word says, I have
to look in my grammar book. With the help of the book I can
decide if I should follow Word or not.
-I do not understand what Word says. I have to ask the teacher
or some other competent person for help.